[tjliang]

Originally Posted by will.logie

Hey Carlos,

I've just had a quick look at my makeAxialMesh example and found that this problem you mention persists for me too when I use makeAxialMesh, regardless of binary and writePrecision.

The example where ascii -> binary worked for me (post from April 6th) does not use makeAxialMesh. There I create the wedge in gmsh (translate mesh -1° about symmetryPlane and extrude/revolve 2° about symmetryPlane) and convert that to foam. So I'm not sure I can help you much further with makeAxialMesh - are you using gmsh?

I'm also not sure if you need to straddle the xy-plane equally or not. Maybe someone else can comment on the interpretation from the user manual:

Hello,

by the way i want to ask you about how you can do that in gmsh. Because when i managed to generate structured mesh in wedge in gmsh with one layer in the rotation direction, all that seems perfect . But after that when i define physical surfaces and volumes, I tried to generate the mesh again , all the meshes turned back to unstructured and in the rotation direction there is not the one layer grid any more. And i looked into the .geo file and fund out that all the codes related with Extrude was disappeared. Sofar i have got no idea about why.

Could you please tell me how to make wedge mesh in gmsh after define physical surfaces?

Bests,

Peng

[will.logie]

Originally Posted by nimasam

another suggestion which works for me was like that:

 fvScalarMatrix p_rghEqn
        (
            fvm::laplacian(rAUf, p_rgh) == fvc::div(phiHbyA)
        );

          p_rgh = p-rhok*gh;
          p_rghEqn.setReference(pRefCell, getRefCellValue(p_rgh, pRefCell));

        p_rghEqn.solve(mesh.solver(p_rgh.select(pimple.finalInnerIter())));

This got me thinking: do the situations we're talking about here (closedVolume) actually need a reference pressure?

Although we're defining one in fvSolution with pRefCell/pRefPoint and pRefValue, isn't it superfluous?


I've been running some simulations with

p_rghEqn.setReference(pRefCell, getRefCellValue(p_rgh, pRefCell));

commented out and so far like what I've been seeing.

[mikulo]

Originally Posted by Yahoo

Hi
I have developed a solver which works fine on single processor, but diverges when I am running it on multiple processors. For the later, by changing the cell at which pressure is set (pRefCell), my results change. i.e. pRefCell is affecting the results on parallel simulations! Any hints?

Hello, i have encountered same problem with you too. I knew this is too late, but have you found a solution for this problem? Please help..